This invention relates to a method of recording digital audio and video signals on-a magnetic tape, and to a digital video tape recorder for recording signals according to this method and playing back the recorded signals.
Video tape recorders, also known as video cassette recorders, are widely used both in the television broadcast industry and at the consumer level for recording and playing back video signals with associated audio signals. The audio and video signals are recorded on a magnetic tape in a series of helical tracks oriented at an angle to the long axis of the tape. Recently digital video tape recorders, which offer superior noise performance and editing capabilities, have come into use. In recording, a digital video tape recorder samples analog audio and video input signals, encodes the sample values digitally, and records the digital data on the tape. In playback, the digital video tape recorder reads digital data from the tape, decodes the data, and generates analog audio and video output signals.
The prior art of recording digital audio and video signals is embodied in, for example, the D-2 format developed by the Society of Motion Picture and Television Engineers, in which each helical track comprises a video sector and two short audio sectors, the audio sectors being disposed at the two ends of the track. The audio sectors are thus located near the edges of the tape, which are the areas most prone to burst errors caused by scratches. Various error countermeasures are taken. For example, an error-correcting code is recorded together with the audio and video data, permitting the correction of burst errors up to a certain length. Also, the audio signal, which normally comprises four channels, is recorded with 100% redundancy, identical data being written in the audio sectors at the upper and lower edges of the tape.
The video signal is divided into frames, a frame corresponding to one complete image on the screen. A frame is divided into two fields comprising the even and odd raster lines on the screen, respectively. In the D-2 format, a frame comprises twelve consecutive helical tracks on the tape, each of its constituent fields comprising six consecutive tracks.
The audio signal is recorded in a continuous manner without being divided into frames and fields in any special way on the tape. Most digital video tape recorders, however, internally divide the audio signal into frames at the same points as the video signal and process the audio signal one frame at a time.
When a tape is edited, the editing normally begins and ends at a boundary between video frames, which provides a clean break in the video image. This practice is generally followed even when the editing is audio dubbing, in which the video signal is left unchanged but the audio signal is replaced with a new audio signal. In the prior art, since the same frame boundaries are used for the audio and video signals, after audio dubbing each audio frame consists entirely of the old audio signal or entirely of the new audio signal.
One problem with the D-2 format is that it adopts an inefficient means of coping with burst errors: recording the audio signal with 100% redundancy uses excessive space on the tape. This problem becomes even more pronounced when the D-2 format is adapted to a two-channel audio signal. (Many digital video tape recorders are built to accept both two-channel and four-channel audio signals.) If the same recording parameters (such as sampling frequency and number of bits per sample) are used as for a four-channel signal, the D-2 format requires 200% redundancy for a two-channel signal.
A further problem is that, the question of redundancy aside, the disposition of the audio sectors in narrow strips near the edges of the tape increases their vulnerability to burst errors, while the relative shortness of the audio sectors limits the error-correcting capability of the error-correcting code.
A problem that occurs when audio dubbing is performed using the D-2 format is that, while a clean break may be desirable in the video signal, it causes noise in the audio signal. Irritating clicks are heard at the points of change from the old audio signal to the new audio signal and from the new audio signal to the old audio signal. (These points will be referred to herein as the edit points.) Complex schemes have been used to combat this edit-point noise, but without complete success.